Forbes

Big Data At Tesco: Real Time Analytics At The UK Grocery Retail Giant

By Bernard Marr

December 13, 2016

Tesco is the UK’s largest food retailer and has long been a pioneer when it comes to technology and data. It was one of the first supermarket chains to begin tracking customer activity through its loyalty card system and has successfully managed the transition to online retailing.

Now it is facing up to the challenges brought about by the latest advancements in technology – the quest for real time data analytics, big data and the efficiencies made possible by the emerging Internet of Things.

Last year, it appointed a new CTO – Edmond Mesrobian, formerly CTO of Expedia – who made it clear that all those technologies are key to staking Tesco’s place in the future of grocery retailing.

Applying cutting edge analytics and the most up-to-date data is the supermarket’s answer to dealing with obstacles ranging from evolving customer behavior, to facing up to newer competitors. Many of these (such as Amazon, which recently began delivering fresh groceries) are built from the ground up as digital and data-driven organizations.

Much of the business’s innovation begins within the Tesco Labs division, which was founded to research new technologies which could benefit the supermarket and its customers. Tesco Labs has up to 50 projects on the go at any one time and experiments with VR and AR, connected home devices, near field communications and mobile applications. It also runs regular hackathon events where programmers compete to develop new solutions.

Key challenges

Key challenges for Tesco in the future include:

  • Continuing to gain a better understanding of the changing nature of consumer behavior
  • Creating efficiencies in their logistics and distribution chains, to keep down costs and minimize environmental impact
  • Facing up to the challenge of emerging business models which compete with their own
  • Reducing the amount of food which goes to waste at their stores

The answer to all of these is thought to lie in real time analytics and the growing amount of customer, operational and external data which is available to the business, thanks to the Internet of Things.

Vidya Laxman, Tesco head of global warehouse and analytics, tells me “We are focused on data now and realize that to get where we want to be in five years time, we have to find out what we will need now and create the right infrastructure.

“We are focusing on Teradata and Hadoop in our ecosystem – how do we get both of those working together and doing the right thing – that’s what our strategy is all about. And the outcome of all of this will be to have real time analytics – that’s what we are working towards.”

Predicting sales

One area of operations where data is very much at the forefront is sales forecasting. modeling trends in customer behavior has thrown up some surprising insights. “It can be something like how will people shop in a store over a week?,” says Tesco head of forecasting and analytics, Mike Moss, “And not just how they shop in each store – but how they shop for each product.

“Using analytics and clustering and suchlike, we found that the way we thought products hung together – the way we buy products – is not really the way products behave.”

With 3,500 stores in the UK alone (as well as substantial numbers in overseas markets such as Thailand and India) and each store stocking an average of 40,000 products, tracking them all just once involves the creation of over 100 million data points. This is where in-database analytics comes into play – deploying analytics technology where the data is stored, rather than moving the data in batches for external analytics.

“We can use clustering to make sure products are predictable and behave in the right way, and that means we order them in the right way and they will always be in stock, and not going to waste,” Mike says.

The Tesco Data Lake

Tesco is in the process of moving to a data lake model, based around the Hadoop framework. This will be a centralized, cloud based repository for all of its data, codified in a way to make it accessible and useable by any arm of the operation, as and when it is needed.

Laxman says “Different stakeholders need different data – finance is a stakeholder and they need sales information and forecasts. Supply chain is another one, then you have the customer side focused on marketing and campaigns. We have data scientists in all of our organizations who need access to the data.

“That’s where Hadoop comes into the picture. We’ve just started on this journey – we’ve had data warehousing for some time so there are some legacy systems present and we want to leverage what’s good and see where we can convert to using new strategies.”

Getting access to data needed is often the biggest problem when a new strategy is devised, Laxman explains – “What happens is it takes at least 9 to 10 months just to ingest the data, under the traditional way of working. So we are working on changing that so the speed is much higher – we are going through quite a few transformations not only from a technology aspect but also in methodology.”

Sensing the future

Sensor data is also increasingly being used across the organization. One implementation involves monitoring the temperature of fridges and freezers across the network of stores. Every machine is monitored centrally, and predictive algorithms are used to determine when a particular unit will need servicing, or when replacement parts will need to be ordered.

The business is also rapidly becoming more involved in open source development. “It’s a big game changer,” Moss tells me.

“When I think back to eight years ago when I build my first forecasting system [at Tesco] I used open source and I had to go through so much to get it signed off. There wasn’t the trust there in the software.

“It now feels like we’re in a very different place than previously – when you might only have been allowed on a Microsoft stack, for example.”

Developers, engineers and data scientists at the company are now encouraged to use open source technology wherever possible, and also to put back into the OS communities where the technologies are evolved.

“Now we have freedom and all the engineers can use what they need to use, as long as its reasonable and it makes sense. We use Github to push code back into the community. It’s about growing the pie rather than trying to keep your own slice of it.”

So it’s clear that despite the huge shift in shopper and competitor behavior since e-commerce went mainstream, Tesco hasn’t stood still. By keeping on the cutting edge of technological advancement, it hopes to retain its market share against seemingly more agile and technologically-driven upstarts. Combining the benefits of its existing systems – its worldwide network of stores and its distribution infrastructure – with the possibilities brought about by new technology including big data and real time analytics, is clearly seen as the recipe for success.

 

This article was written by Bernard Marr from Forbes and was legally licensed through the NewsCred publisher network.